On this concept: Formula

Not sure which lesson to do first? Find your level and get a plan →

Math · Advanced Functions · Grade 9-12

Sensitivity

Sensitivity measures how strongly the output responds to a small change in the input: high sensitivity means a tiny input change causes a large output change.

Sensitivity≈ΔfΔx=f(x+Δx)−f(x)Δx

Quick Answer

Sensitivity measures how strongly the output responds to a small change in the input: high sensitivity means a tiny input change causes a large output change. You estimate it with ΔfΔx: nudge the input a little, see how far the output moves, and divide. Use it to find where a measurement error matters most, which input in a formula deserves the most care, or where a curve is steep. For most functions the answer depends on where you look: f(x)=x2 barely responds near 0 and responds strongly near 10.

Why This Matters

Sensitivity tells you where precision matters. A square tile's area is s2, and a 1 mm error in the side changes the area about ten times as much for a 1 m slab as for a 10 cm tile. Engineers, nurses calculating doses, and anyone fitting a model use it to decide which inputs to measure carefully. Without it you treat every input error as equally harmless.

Intuitive Explanation

Take f(x)=x2 and nudge the input by 0.1 at a few places. From 1 to 1.1 the output goes from 1 to 1.21, a change of 0.21, so the ratio ΔfΔx is 2.1. From 5 to 5.1 it goes from 25 to 26.01, a change of 1.01, ratio 10.1. The same small nudge moves the output almost five times as much at x=5. Sensitivity is that ratio: output change per unit of input change.

On a graph, sensitivity is steepness. A straight line has the same steepness everywhere, so g(x)=3x is equally sensitive at every input: any nudge of 0.1 moves the output 0.3. Curves change steepness, so their sensitivity is local, and you must say where you measured it. As Δx shrinks, the ratio settles on a single number, the derivative f′(x), the exact sensitivity at that point.

A large output is not the same as a sensitive one. The function h(x)=1000 has huge outputs and zero sensitivity, since no nudge changes it. Meanwhile p(x)=50x near x=0 has outputs close to 0, yet every nudge of 0.1 moves it by 5. What matters is how far the output moves, not how big it is. Report the size of your nudge as well: on a curve, a ratio measured over Δx=0.1 and one over Δx=2 can differ a lot.

Think of it like this. A coarse knob and a touchy one: on the touchy knob a tiny twist makes a big jump in loudness, while on the coarse one you can turn a long way and hear little change. A curved function is a knob whose touchiness changes as you turn it, gentle at one end and sharp at the other.

Core idea. Sensitivity measures how much the output moves for a tiny change in the input — steep means very sensitive.


When to Use

Use sensitivity when a problem asks how much the output changes for a small change in the input, which input matters most, or how an error in a measurement spreads into a result. Compute f(x+Δx)−f(x)Δx with a small Δx at the point in question. If the question is whether small disturbances die out over many repeated rounds, use stability instead; if the function is a straight line, its slope already gives the sensitivity everywhere.

Tip. Ask: Does a small change in the input produce a large change in the output here?

How to Recognize It

Before measuring sensitivity, fix the input value you are looking at and the size of the nudge, because one function can be touchy in one place and calm in another.

  1. Is the function a straight line?

    If yes, its slope is the sensitivity at every input, and there is nothing local to work out. If it curves, compute f(x+Δx)−f(x)Δx at the particular input the question names.

  2. Does the question compare inputs measured in different units, such as metres and degrees?

    If yes, compare percent changes instead of raw slopes: ask what a 1% change in each input does to the output. A slope per metre cannot be ranked against a slope per degree.

  3. Does the question want the exact responsiveness at one point, rather than an estimate from a small step?

    If yes, shrink Δx toward 0; that limit is the derivative. A finite step gives a good estimate when the curve is nearly straight over that step.

  4. Is the question about whether a disturbance dies out over many repeated rounds?

    If yes, that is stability, which follows a disturbance through a loop. Sensitivity is about a single reaction: one input change, one output change.

Sensitivity compared with Slope of a line, Derivative (instantaneous sensitivity), and Stability

All four rows measure how one quantity responds to another, so they are easy to treat as one idea. Ask whether the response is the same everywhere, measured at one exact point, or followed through many repeated rounds.

Sensitivity

Meaning
Use this when you want to know how much the output changes for a small change in the input near a point.
Key test
Does a small change in the input produce a large change in the output here?
Formula
Sensitivity≈ΔfΔx=f(x+Δx)−f(x)Δx
Example
For f(x)=x2, is the output more sensitive to a small input change near x=1 or near x=5?

Slope of a line

Meaning
Constant rate of change for a straight line, the same everywhere.
Key test
Use when the relationship is linear so sensitivity is one fixed number.
Formula
m=ΔyΔx
Example
y=3x moves 3 per unit, everywhere

Derivative (instantaneous sensitivity)

Meaning
The exact sensitivity at a single point, the limit as Δx→0.
Key test
Use when you need precise local responsiveness, not an average over an interval.
Formula
f′(x)=dfdx
Example
Speedometer reading at one instant

Stability

Meaning
Whether disturbances die out over repeated steps, not the size of one response.
Key test
Use for long-term behavior of a feedback loop, not a single input-output reaction.
Formula
∣f′(x∗)∣<1
Example
Ball returning to a bowl

Formula & Notation

Sensitivity≈ΔfΔx=f(x+Δx)−f(x)Δx
S(x)=f′(x)=lim⁡Δx→0f(x+Δx)−f(x)Δx; relative sensitivity =xf(x)⋅f′(x)

How to read it: ΔfΔx denotes the average sensitivity. dfdx or f′(x) denotes the instantaneous sensitivity (derivative).

Worked Examples

Example 1 — Compare two points

Easy

For f(x)=x2, is the output more sensitive to a small input change near x=1 or near x=5?
Show the steps and the answer
  1. Nudge near x=1: f(1.1)−f(1)=1.21−1=0.21, so ΔfΔx=0.210.1=2.1.

    Using the same small Δx at both points makes the two ratios fair to compare.

  2. Nudge near x=5: f(5.1)−f(5)=26.01−25=1.01, so ΔfΔx=1.010.1=10.1.

    The output moves almost five times as far for the same nudge.

  3. Compare: 10.1>2.1.

    The parabola is steeper at x=5, so the output responds more strongly there.

  4. Shrink the nudge to check: f(5.01)−f(5)0.01=25.1001−250.01=10.01.

    A smaller nudge gives nearly the same value, so the conclusion does not depend on choosing 0.1.

Answer: More sensitive near x=5

Takeaway: Compare sensitivities using the same small nudge at each point; on a curve, the steeper place is the more sensitive one.

Example 2 — Flat region

Standard

For the same f(x)=x2, how sensitive is the output near x=0?
Show the steps and the answer
  1. Nudge from 0 to 0.1: f(0.1)−f(0)=0.01−0=0.01.

    Near the vertex the parabola is almost flat, so the output barely moves.

  2. Divide: 0.010.1=0.1.

    Sensitivity is the output change per unit of input change.

  3. Compare with x=5, where the same nudge gave 10.1.

    The function is about 100 times as sensitive at 5 as at 0, so a sensitivity is meaningless without a location.

  4. Shrink the nudge: with Δx=0.01 the ratio is 0.00010.01=0.01.

    At a flat point the ratio keeps shrinking with the nudge, toward the exact value 0.

Answer: Very low, ≈0.1

Takeaway: Sensitivity is local: one function can be nearly unresponsive at one input and very responsive at another.

Example 3 — Find the mistake: big output, small response

Application

Compare g(x)=500+x and h(x)=10x near x=2. Omar writes: "g(2)=502 is much bigger than h(2)=20, so g is more sensitive near x=2." Find the mistake and correct it.
Show the steps and the answer
  1. Nudge g: g(2.1)−g(2)=502.1−502=0.1, so ΔgΔx=0.10.1=1.

    Sensitivity measures how far the output moves, not where it starts.

  2. Nudge h: h(2.1)−h(2)=21−20=1, so ΔhΔx=10.1=10.

    Every 0.1 of input moves h by a full unit.

  3. Name the error: Omar compared output sizes, 502 and 20, instead of output changes.

    The 500 in g only lifts the whole graph; it adds nothing to its steepness.

Answer: h is more sensitive: about 10 output units per input unit, against 1 for g.

Takeaway: Judge sensitivity by the change in output per change in input; a large constant makes outputs big without making them responsive.

Where You'll See It

  • Measuring tiles and slabs

    A square's area is A=s2, so measuring the side 1 mm too long adds (s+1)2−s2=2s+1 square millimetres. For a 100 mm tile that is 201 mm²; for a 1,000 mm slab it is 2,001 mm². The bigger the square, the more the same measuring error matters.

  • Medicine doses by body weight

    Children's doses are often set per kilogram, for example 15 mg per kg. The dose then has a sensitivity of 15 mg per kg, so misjudging a child's weight by 2 kg changes the dose by 30 mg. That is why children are weighed rather than guessed at.

Common Mistakes

  1. The slip: Calling a function 'sensitive' globally

    Instead: sensitivity is local and varies with where you measure it.

  2. The slip: Confusing a large output value with high sensitivity

    Instead: what matters is the change in output per small input change, not the output's size.

  3. The slip: Ignoring the input scale

    Instead: sensitivity is a ratio ΔfΔx, so the size of Δx matters when reporting it.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. For f(x)=x3, estimate the sensitivity at x=2 using Δx=0.1.

    Hint: Compute f(2.1)−f(2)0.1.

  2. For g(x)=5x+3, find the sensitivity near x=1 and near x=100.

    Hint: Find how much g changes when x goes up by 1 in each place.

  3. A circle's area is A=πr2. The radius changes from 10 cm to 10.1 cm. Find the change in area and the sensitivity.

    Hint: Find π(10.12−102), then divide by Δr.

  4. For h(x)=x, compare the sensitivity near x=1 and near x=100, using Δx=1 in each place.

    Hint: Compute 2−1 and 101−100.

  5. For f(x)=4x−7, do you need to say where you measure the sensitivity, as you would for y=x2?

    Hint: Does the steepness of this graph change anywhere?

  6. A student says 'f(x)=1000+0.1x is more sensitive than g(x)=2x near x=10, because 1001 is much bigger than 20.' Find and correct the mistake.

    Hint: Sensitivity is about change in output per change in input.

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

Why is x so sensitive near 0?

The square root rises steeply at the start: 0=0\sqrt{0} = 0 but 0.01=0.1\sqrt{0.01} = 0.1, so an input change of 0.01 moves the output ten times as far. Near x=100x = 100 the same nudge barely registers: 100.01≈10.0005\sqrt{100.01} \approx 10.0005. The ratio ΔfΔx\frac{\Delta f}{\Delta x} is about 10 in the first case and 0.05 in the second. So the error in a computed square root depends heavily on where the measurement lies.

What is the butterfly effect?

It is a name for extreme sensitivity in systems such as the weather, where a tiny difference in today's conditions can grow into a completely different forecast weeks later. The meteorologist Edward Lorenz found it in 1961 when he re-entered a number rounded to three decimal places instead of six, and his computer weather model drifted to a different result. Small errors there are multiplied over and over, which is why detailed forecasts lose their value after about two weeks.

Why does a 1% error in a ball's radius cause about a 3% error in its volume?

The volume is V=43πr3V = \frac{4}{3}\pi r^3, so multiplying rr by 1.01 multiplies VV by 1.013≈1.03031.01^3 \approx 1.0303, about 3% more. In general, when the output is a power of the input, the percent error is multiplied by the exponent: squaring doubles it, cubing triples it, and a square root halves it. So if a volume must be known to within 3%, the radius must be measured to within 1%.

Is sensitivity the same as how steep the graph looks?

Only when both axes use the same scale. Steepness on paper can mislead: stretch the yy-axis and every curve looks steeper, while the sensitivity, a ratio with units such as dollars per hour, stays exactly the same. Two graphs of one function drawn with different axis scales can look completely different. Read the numbers off the axes and compute ΔfΔx\frac{\Delta f}{\Delta x} rather than trusting how tilted the curve appears.

Learning Path

← Before

Rate of Change
Sensitivity

You are here

Next →

Derivative
You need Rate of Change: the ratio ΔyΔx between two points. Sensitivity aims that ratio at a question, how much a small input change matters at one particular place, and shows that on a curve the answer changes from point to point. Next, the Derivative takes the limit as Δx→0, turning this estimate into the exact sensitivity f′(x) at a single input.

See Also